JP2008084126A - Character correction confirmation method in character recognition system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve such a problem that characters which can be input are determined by the kind of character input to a field, and consequently it takes much time in switching fields. <P>SOLUTION: A mode (a field mode) is prepared in which, when performing a correction job of the same form, only a common field (for example, 'address') to a plurality of pages is corrected, the next field is corrected after the previous correction, and this is repeated. A page mode/field mode can be switched for each form. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  This patent relates to a character recognition system that receives an electronic image, executes form recognition and character recognition, and can perform a character correction operation.

  Conventionally, application forms, application forms, etc., are sent as paper media to the office management center of the company and stored in a database within the company via human resources. In recent years, paper media has changed to fax images and electronic images read by scanners, and has shifted to work on personal computers. It has become a movement to automatically perform work by receiving electronic images and automatically performing form recognition, character recognition, etc.

As a conventional example, for example, at the stage of correcting the character recognition result, it is possible to select whether the correction is performed in the form unit / field unit / character unit, and the result is displayed instead of the character recognition result. There is a configuration in which converted characters are collectively converted into converted characters (see Patent Document 1).
JP-A-9-114918

  In recent years, it has become possible to automatically process received data by introducing business software. However, for example, form recognition and character recognition cannot be said to be complete recognition accuracy, and work such as manual correction is required. It has occurred. Here, when correcting a character, each field arranged in the page is corrected in a page unit, and after the completion of one page, the next page is corrected. The following problems occurred in this configuration.

  (1) The characters that can be entered are determined by the type of characters entered in the field, and it takes time to switch the field.

  (2) I don't know the remaining work time.

  (3) One worker can look over the entire form, which causes a security problem.

  In the present system, the following configuration is adopted in order to eliminate the above-described drawbacks.

  (1) Revision of recognition results was done in field units instead of page units.

  (2) The configuration is such that the number of remaining jobs can be displayed on the screen for modifying the application form.

  (3) The configuration is such that the number of workers currently working can be displayed on the screen for modifying the application form.

  (4) The configuration is such that the remaining time of work can be displayed on the screen for modifying the application form.

  (5) When only work in a certain field has advanced in advance, the person in charge of the remaining work can be notified by e-mail, facsimile, telephone, etc.

  (6) When only the work in a certain field has advanced in advance, the work cannot be done so that the work does not proceed further.

  (7) In the above case, the guarded operation can be performed again when the difference is reached.

As mentioned above
(1) The recognition result correction work was done in field units instead of page units.

  (2) The configuration is such that the number of remaining jobs can be displayed on the screen for modifying the application form.

  (3) The configuration is such that the number of workers currently working can be displayed on the screen for modifying the application form.

  (4) The configuration is such that the remaining work time can be displayed on the screen for modifying the application form.

  (5) When only work in a certain field has advanced in advance, the person in charge of the remaining work can be notified by e-mail, facsimile, telephone, etc.

  (6) When only the work in a certain field has advanced in advance, the work can not be done so that the work does not proceed further.

(7) In the above case, the guarded operation can be performed again when the difference is reached.
With that
-It has become possible to remove the trouble of switching fields.

  -Made the worker aware of the remaining work time.

  ・ Security has been protected.

  Below, with reference to FIGS. 1-14, one Embodiment of the apparatus which concerns on this invention is described. FIG. 1 is a block diagram of a system according to the present invention.

  Reference numerals 101 and 102 denote client PCs, in which a system administrator registers a new data file, performs system maintenance processing such as correction of user management information, or a registration worker confirms sorting of electronic documents received on a server. Or used by an approver to perform approval work. A scanner 103 reads image data shared on the network, and reads an application form sent by mail from a user, a document brought into a branch, or the like.

  Reference numeral 104 denotes a multifunction device shared on the network, which can be used as a scanner as in the case of 103. Further, the FAX image received by the 104 can be registered in a 109 server or the like via a LAN.

  Reference numerals 105 and 111 denote local area networks (LANs), and the devices 101 to 105 connected to the network exchange data with other devices via the LAN. Reference numeral 107 denotes the Internet, and reference numeral 106 denotes a client PC directly connected to the Internet. Reference numeral 108 denotes an HTTP server that receives a request that can be HTTP protocol from a client PC via a network. Reference numeral 109 denotes a Web application server. Several web application servers are registered in the HTTP server, and the HTTP server allocates processing to the appropriate web application server according to the request from the client PC. In addition, the Web application server that has received the request executes processing and returns the processing result to the client. A database 110 records information such as received data, character recognition results, application processing status, and system information such as user / group information and registration data file processing information.

  As described above, the WEB database system functions organically in combination with the HTTP server 108, the WEB application server 109, and the database 110.

  Next, FIG. 2 is a block diagram illustrating the configuration of the apparatus according to the present invention.

  Reference numeral 1 denotes a system bus, and each component block described below is connected to the system bus. Reference numeral 2 denotes a CPU (Central Processing Unit). Reference numeral 3 denotes a program memory (referred to as PMEM). A program for this processing is appropriately selected / read from the hard disk 10 and executed by the CPU 2. Data input from the keyboard 12 is stored as code information in the PMEM which is also a text memory.

  A communication control unit 4 controls input / output data in the communication port 5. A signal output from the communication port 5 is transmitted to a communication port of another device on the network via the communication line 6.

  In this embodiment, a network such as a LAN is described, but it goes without saying that the present invention is applicable even if the communication port and communication line connected to the communication control unit are general public lines.

  8 is an external storage device control unit, 9 and 10 are data file disks, for example, 9 is a flexible disk (referred to as FD), and 10 is a hard disk (referred to as HD).

  Reference numeral 11 denotes an input control unit to which input devices such as 12 keyboards and 13 mice are connected. The operator operates the keyboard 11 to give a system operation command or the like.

  Reference numeral 13 denotes a pointing device (referred to as PD) for instructing processing of image information on 16 CRTs. In this embodiment, a mouse is used. As a result, the cursor on the CRT 16 is arbitrarily moved in the X and Y directions to select a command / icon on the command menu to instruct processing, in addition to instructing an editing target and drawing position. 14 is a video image memory (referred to as VRAM), 15 is a display output control unit, and 16 is a CRT. Data displayed on the 16 CRTs are expanded as bitmap data on the 11 VRAMs.

  Reference numeral 17 denotes a printer control unit that controls data output to the connected printer 18.

  Reference numeral 1A denotes an image reading device control unit that controls the connected image reading device 1B.

  1 has the same function even if the image reading device and the image reading device are physically separate components, even if the image reading device is a single component including the image reading device. And

  Note that the program stored in the ROM in this embodiment may be stored in a storage medium such as a hard disk (HD) or a flexible disk (FD) directly connected to the apparatus. Furthermore, it may be stored on another device connected via a network. The program of the present invention can be supplied to a system or apparatus via a storage medium such as FD or HD or a network.

FIG. 3 is a block diagram illustrating a module configuration according to the present invention.
A client PC 301 stores an information registration module 302 for registering data files such as catalog information and image data in the server, and an output form registration module 303 for registering a form for outputting product data. However, since these modules are automatically distributed from the server PC in the plug-in format of the web browser as needed, there is no need to perform installation processing on the client PC.

  Reference numeral 304 denotes a network for exchanging data, and reference numeral 305 denotes an HTTP server.

  Further, reference numeral 306 denotes a Web application server, which stores a module 307 for processing a request from the client PC received from the HTTP server. Corresponding functions include user authentication, search processing, received data registration, character recognition result registration, form maintenance function, and the like. These modules are loaded on the memory in response to a request from the client and execute processing.

  In addition to these modules, there are editing modules 308 such as a workflow editor section and a cover sheet editor section. Further, there are a library 309 and various utility libraries 310 for exchanging data with the database 311.

  FIG. 4 is a block diagram for explaining a processing flow according to the present invention. In FIG. 4, S1 indicates a process of fetching received data from a multifunction device such as a FAX, a scanner, or a program mountable multifunction device via a LAN. S2 shows a step of performing form recognition (FR) of the received data and determining the form type. S3 indicates a step of performing character recognition (CR) and bar code recognition (BR) and extracting characters described in the form found in S2. S4 is a step where a person in charge called a post editor (PE) modifies the characters recognized in S3 while viewing the PC screen, S5 is a step for realizing an application approval workflow (PA flow), and S6 is received by the image server. A step of registering data, S7 is a step of registering data in a back-end server such as a company, and S8 is a sending letter such as registration confirmation to a sending user such as an application form that is received data. This is the step of sending by mail.

  In the present invention, the control of FIG. 4 is called a workflow or a business process, and each step from S2 to S8 can be switched ON and OFF. A business process is assigned to each received application.

  FIG. 12 is an example for explaining a conventional method for correcting a form recognition result. Hereinafter, a method for correcting character recognition will be described with reference to FIG. 1201 is a display example on the CRT of the personal computer. Reference numeral 1202 denotes a received image. Reference numeral 1203 denotes a character written by handwriting. "Yamada" is written in the reading field, "Yamada" is written in the name field, and "00-1234-5678" is written in the Tel field. 1204 is a work ID assigned by the system. This form is set to “000001”. 1205 is a title display of “field”, and 1206 is a title display of a recognition result. Reference numeral 1207 denotes the recognition result “Yamada”. This character string displays the result of recognizing the reading part from 1202 images. Similarly, “Yamada” 1208 is displayed as a recognition result, and “00-1234-5678” is displayed as a recognition result 1209. Hyperlinks are made from 1207 to 1209, and the display changes to FIG. 13 when 1208 is clicked on with a mouse. In FIG. 13, reference numeral 1301 denotes a display screen example. Reference numeral 1302 denotes the title of the field name, and the field name “name” clicked in FIG. 12 is displayed. Reference numeral 1303 denotes an area image on the electronic image. 1203 copies of the enlarged image are displayed. Reference numeral 1304 denotes a content correction area. Initially, the character “Yamada” displayed at 1209 is described in 1304, but it can be freely modified thereafter, and a character string such as “Taro Yamada” as described in 1304 can also be described. . Reference numeral 1305 denotes a completion button. When the user presses 1305 from the example of FIG. 13, the processing returns to FIG. 12 and 1208 is “Taro Yamada”. That is, the corrected result overwrites the recognition result. Reference numeral 1306 denotes a next button. In this example, when the 1306 button is pressed, the recognition result in the “Tel” field 1209 is displayed. As described above, the recognition result correction method is configured to sequentially correct all fields in one page.

  FIG. 5 is a flowchart diagrammatically illustrating the method shown in FIG. 12 and the operation of the present invention. This will be described with reference to FIG. FIG. 5 is an example diagram for explaining the order in which three identical application forms are corrected. In the conventional operation shown in FIG. 12, the work proceeds for each application, and worker A corrects all the forms of one page with the blue field, the yellow field, and the red field. When there are a plurality of workers, the same work as A can be performed like worker B and worker C. However, it takes time to input each time the field to be corrected changes in the conventional example. In other words, in the blue "Name" field, you can correct hiragana and kanji, but in the yellow "phone number" field, you need to change the keyboard input method or type because it is only a correction of numbers. The keys are clearly different. There is also a problem that one operator can see all the forms. Since all fields can be confirmed, that is, the information related to privacy, such as where and where the annual income is, can be confirmed at a glance. The operation of the present invention will now be described in the lower part of FIG. In the present invention, the same field of a plurality of received forms is corrected and confirmed. That is, the worker A is determined in advance as a worker in the blue “name” field, and the “name” field of the reception form is corrected and confirmed. Similarly, worker B confirms the telephone number field and worker C confirms the address field. As a result, the work efficiency can be improved by simple work of only the same field, and privacy can be protected.

  FIG. 6 is an operation example of the present invention. Only the telephone number field in FIG. 5 is displayed for correction. Reference numeral 61 denotes an area for displaying an image. 62 is a page on which correction work has been completed, 63 is a page being corrected, and 64 is a telephone number field included in the next form. 65 is a field where 63 recognition results are first input, and a text input area for the operator to correct the recognition results while viewing 63 images. 66 is a confirmation key. This button is pressed. Reference numeral 67 denotes a return key, which is a key for exiting from the correction work. If there is a next form as in this example, the display continues to shift each time the enter key is pressed.

  FIG. 7 is a display example of the present invention. In FIG. 7, reference numeral 71 denotes an area where the number of remaining jobs is displayed.

  A counter that is subtracted each time the enter key is pressed and added if there is a received image is displayed. This example shows that there are 10 unconfirmed forms including the target field. Reference numeral 72 denotes the number of workers currently participating. This example shows that five people are editing the yellow area. 73 indicates how long the correction work for this target field will be completed. In this example, it ends after 10 minutes. This time may be automatically measured from the number of workers, the speed, and the number of remaining jobs.

  FIG. 8 is an example of a flow when the display shown in FIG. 7 is performed. A description will be given with reference to FIG. This flow starts when the person in charge logs in. 81 is a step in which the system confirms the role when logging in. 82 is a step of ascertaining which field can be modified as a result of 81. 83 is a step of confirming whether or not there is a remaining field to be allocated. If there are any, proceed to 84. 84 is a step of extracting field information from the form reception queue. 85 is a step of displaying 71. 86 is a step for displaying 72, and 87 is a step for displaying 73. Reference numeral 88 denotes a processing time measurement start process serving as reference information 73, and 89 is a step of confirming or correcting 65 text while viewing 63. Reference numeral 810 denotes a step for ending the measurement of the processing time, and reference numeral 811 measures the time taken from 88 to 810, which is used as the work time per worker field. 812 is a step of assigning a job to log-in workers after completion of the work in sequence, and calculating an expected end time assuming that each person in charge performs work with the processing time measured for each worker.

  FIG. 9 is a diagram for explaining the logic of distribution according to the present invention. In this system, there is an application reception queue, and forms that are received and are being worked on are registered. When there is a reception, it will be connected to the tail. Moreover, when an operator logs in, it is allocated from the head sequentially. As a result, the work is leveled. When work for one form is completed by workers in all fields, the work is moved to the field processed queue.

  FIG. 10 is an explanatory diagram of a mechanism at the time of distribution according to the present invention. If correction of only a specific field progresses, there will be a problem that it takes a long time to go to the post-correction process such as an approval process, and a problem that the system is difficult to operate because the memory cannot be released by holding the form. Therefore, when work on a specific field precedes, the mail address of the person in charge of the unprocessed field is automatically acquired and notified to the person in charge by email. In addition, an upper limit counter is provided when work on a specific field precedes, so that correction work on the corresponding field cannot be performed. This operation guard can be notified to the administrator by e-mail. In addition, when the number of processed and unprocessed fields exceeds a certain level, a message that prompts the operator to increase the speed of work is output and displayed.

  FIG. 11 is an explanatory diagram showing the mechanism described in FIG. 10 in a flow. The counter of FIG. 14 is used. A step of determining a field A to be corrected 1101 after login. 1102 is a process for determining whether or not the value obtained by subtracting the processed number (b) of field B from the processed number (a) of field A is equal to the work interruption counter (m). If they are equal, the process proceeds to 1114 to perform the allocation failure process. During this process, e-mail is sent to the administrator. 1103 is a process for determining whether or not a value obtained by subtracting the processed number (c) of field C from a is equal to m. If they are equal, the process proceeds to 1114 to perform the allocation failure process. If not, go to 1104. A process 1104 counts up a. Reference numeral 1105 denotes processing for determining whether or not a−b is equal to the mail transmission counter (x). If the mail transmission counter has been reached, the process proceeds to 1110. Reference numeral 1110 denotes a step of sending an e-mail to an operator having the authority of the field B. Step 1106 is a step for determining whether or not a-c is equal to x. If they are equal to each other, the process proceeds to 1111, and mail is transmitted to the worker having the authority of the field C. Step 1107 is a step for determining whether or not a−b is equal to the display counter (y). If they are equal, proceed to 1112 and send a mail to the person in charge in the field B. 1108 is a step of determining whether a-c is equal to y. If they are equal, proceed to 1113. Reference numeral 1113 denotes a step of sending a mail to the worker in the field C. Reference numeral 1109 denotes a process when the assignment of the field A is successful, for example, a step in which correction work is performed.

1 is a system configuration diagram showing a schematic configuration of an embodiment of the present invention. It is a block diagram explaining the apparatus structure of one Example of this invention. It is a block diagram explaining the module structure of one Example of this invention. It is a block diagram explaining the processing flow of one Example of this invention. Flow diagram illustrating the operation of the present invention. Example of operation of the present invention Display example of the present invention Example flow when performing the display shown in FIG. The figure for demonstrating the logic of distribution by this invention Explanatory drawing of mechanism at the time of distribution according to the present invention An explanatory diagram showing the mechanism described in FIG. Example diagram explaining how to correct the recognition result of a conventional form Operation diagram for field correction according to the present invention Illustration for explaining the counter

Claims (7)

In a system having means for recognizing an electronic image in a form having a plurality of input fields, means for recognizing characters in a field in the form, and means for correcting the recognized result,
A character recognition system that can switch between correction work in page units and field units. Means for receiving an electronic image; means for recognizing an electronic image in a form having a plurality of input fields; means for recognizing characters in a field in the form; means for correcting a recognized field result; and a plurality of operations In a system having means for assigning a field to be corrected for each person, means for displaying the work to be corrected on the screen, and means for performing the correction work on a field basis,
A character recognition system that can display the number of remaining jobs on the correction screen. Means for receiving an electronic image; means for recognizing an electronic image in a form having a plurality of input fields; means for recognizing characters in a field in the form; means for correcting a recognized field result; and a plurality of operations In a system having means for assigning a field to be corrected for each person, means for displaying the work to be corrected on the screen, and means for performing the correction work on a field basis,
A character recognition system that can display the number of workers in the same field on the correction screen. Means for receiving an electronic image; means for recognizing an electronic image in a form having a plurality of input fields; means for recognizing characters in a field in the form; means for correcting a recognized field result; and a plurality of operations In a system having means for assigning a field to be corrected for each person, means for displaying the work to be corrected on the screen, and means for performing the correction work on a field basis,
A character recognition system that can display the expected end time on the correction screen. Means for receiving an electronic image; means for recognizing an electronic image in a form having a plurality of input fields; means for recognizing characters in a field in the form; means for correcting a recognized field result; and a plurality of operations In a system having a means for assigning a field to be corrected for each person and a means for performing correction work on a field basis,
After receiving multiple forms, proceed with the correction of the first field for multiple received forms.
When the number of work completions in the first field in the same form exceeds the number of work completions in any of the other fields by a specified value or more, the person in charge of the unprocessed field is notified of information prompting the work by communication means. Character recognition system characterized by that. Means for receiving an electronic image; means for recognizing an electronic image in a form having a plurality of input fields; means for recognizing characters in a field in the form; means for correcting a recognized field result; and a plurality of operations In a system having a means for assigning a field to be corrected for each person and a means for performing correction work on a field basis,
After receiving multiple forms, proceed with the correction of the first field for multiple received forms.
When the number of work completions of the first field in the same form exceeds the number of work completions of correction of any other field by more than a specified value, guarding is performed so that the subsequent correction of the first field cannot be performed. A character recognition system. 7. The character recognition system according to claim 6, wherein the correction of the first field can be started again when the difference between the number of work completions in the first field and the number of work completions in other fields falls below a specified value.

Images